feat(model): LLM 多供应商适配（Adapter/OpenAI 非流式+SSE 流式/per-role 路由/主备降级/多模态检测/调用记录）

internal/runtime/model/model.go

@@ -0,0 +1,137 @@
+// Package model provides a vendor-agnostic LLM adapter layer for Cyrene v2.
+//
+// It defines a normalized Message/Request/Response vocabulary, a single
+// Adapter interface implemented by OpenAIChatAdapter (which speaks the
+// OpenAI-compatible /chat/completions protocol used by OpenAI, DashScope/百炼,
+// local gateways and proxies), a per-role Router, an automatic fallback Chain,
+// multimodal (vision) detection, and per-call CallRecord accounting.
+//
+// The full tool loop (schema registration, argument binding, result re-feed) is
+// deferred to M2; this package only carries interface-level placeholders
+// (ToolSpec) so the wire contract is ready without executing any loop.
+package model
+
+import "time"
+
+// Role identifies who authored a message. The values are vendor-agnostic and
+// translated to each provider's own role vocabulary by the adapter.
+type Role string
+
+const (
+	// RoleSystem is a system/instruction message.
+	RoleSystem Role = "system"
+	// RoleUser is an end-user message.
+	RoleUser Role = "user"
+	// RoleAssistant is an assistant (model) message.
+	RoleAssistant Role = "assistant"
+	// RoleTool is a tool-result message (reserved for the M2 tool loop).
+	RoleTool Role = "tool"
+)
+
+// ContentType describes the kind of payload inside a Content part.
+type ContentType string
+
+const (
+	// ContentTypeText is a plain-text content part.
+	ContentTypeText ContentType = "text"
+	// ContentTypeImage is an image content part.
+	ContentTypeImage ContentType = "image_url"
+)
+
+// Content is a single, possibly multimodal, part of a message.
+type Content struct {
+	// Type distinguishes a text part from an image part.
+	Type ContentType
+	// Text is the text payload when Type == ContentTypeText.
+	Text string
+	// ImageURL is the image source when Type == ContentTypeImage. It may be a
+	// data: URI or an http(s):// URL.
+	ImageURL string
+}
+
+// Message is a single conversation turn, normalized across providers.
+type Message struct {
+	Role Role
+	// Content carries one or more parts. A text-only message uses a single
+	// ContentTypeText part.
+	Content []Content
+	// ToolCallID and ToolCallName are reserved as interface-level placeholders
+	// for the M2 tool loop; they are ignored by current adapters.
+	ToolCallID   string
+	ToolCallName string
+}
+
+// TextMessage builds a text-only message.
+func TextMessage(role Role, text string) Message {
+	return Message{Role: role, Content: []Content{{Type: ContentTypeText, Text: text}}}
+}
+
+// ImageMessage builds a message carrying optional text plus an image. imageURL
+// may be a data: URI or an http(s):// URL. An empty text omits the text part.
+func ImageMessage(role Role, text, imageURL string) Message {
+	parts := make([]Content, 0, 2)
+	if text != "" {
+		parts = append(parts, Content{Type: ContentTypeText, Text: text})
+	}
+	parts = append(parts, Content{Type: ContentTypeImage, ImageURL: imageURL})
+	return Message{Role: role, Content: parts}
+}
+
+// ToolSpec is an interface-level placeholder for a tool/function schema. The
+// full tool loop (registration, argument binding, result re-feed) is M2; here
+// it only lets a request advertise tools to the provider in the wire contract.
+type ToolSpec struct {
+	Name        string
+	Description string
+	// Parameters is the JSON Schema for the tool arguments. Populated from M2.
+	Parameters any
+}
+
+// Request is a completion request, normalized across providers.
+type Request struct {
+	// Model overrides the adapter's configured default model. Empty means the
+	// adapter default (or the chain step model) applies.
+	Model    string
+	Messages []Message
+	// Temperature is the sampling temperature. nil means the provider default.
+	Temperature *float64
+	// MaxTokens caps the generated tokens. Zero means the provider default.
+	MaxTokens int
+	// Purpose labels the call for CallRecord accounting (e.g. "chat").
+	Purpose string
+	// Tools is the interface-level placeholder for M2 function calling.
+	Tools []ToolSpec
+}
+
+// Response is a normalized completion result.
+type Response struct {
+	// Model is the model that produced the response.
+	Model   string
+	Content string
+	// Message is the normalized assistant message.
+	Message Message
+	// PromptTokens, CompletionTokens and TotalTokens report provider usage when
+	// available; zero when the provider does not return it.
+	PromptTokens     int
+	CompletionTokens int
+	TotalTokens      int
+}
+
+// CallRecord describes a single successfully completed LLM call, for the M9
+// monitor/frontend. A failed attempt emits no record.
+type CallRecord struct {
+	// Model is the model that served the request.
+	Model string
+	// Purpose labels what the call was for.
+	Purpose string
+	// Tokens is the total token consumption reported by the provider.
+	Tokens int
+	// TookMs is the end-to-end call latency in milliseconds.
+	TookMs int64
+}
+
+// Options carries per-call knobs layered on top of the Request.
+type Options struct {
+	// Timeout bounds a single HTTP call. Zero uses the adapter client timeout.
+	Timeout time.Duration
+}